From: Keir Fraser Subject: Re: [PATCH 0 of 3] management tools portability Date: Mon, 5 Jun 2006 16:59:09 +0100 Message-ID: References: <1149519637.19111.7.camel@basalt.austin.ibm.com> <1149522607.19111.20.camel@basalt.austin.ibm.com> Mime-Version: 1.0 (Apple Message framework v624) Content-Type: text/plain; charset=US-ASCII; format=flowed Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <1149522607.19111.20.camel@basalt.austin.ibm.com> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: Hollis Blanchard Cc: xen-ppc-devel@lists.xensource.com, Ian Pratt , xen-devel@lists.xensource.com, Christian Limpach List-Id: xen-devel@lists.xenproject.org > I believe the agreement we reached earlier was that *internally* (on > any > side of any interface), passing around a single PFN can be type > 'unsigned long', since on 32-bit systems that still lets you manage 42 > bits of physical memory, and that "should be good enough for anybody." Yes. It's only the interface we want to change. > Unrelated to that, I converted all 'unsigned long' in the *interface* > to > be u64. The one exception is that PFN arrays (not single PFNs) became > 'xen_pfn_t'. > > Does that make sense? It seems weird/arbitrary to me to change the type only of PFNs that are array elements. -- Keir > -- > Hollis Blanchard > IBM Linux Technology Center >